Is "Easy Advanced Control" Really that Easy?
Emerson and some of the other vendors have been promoting "Easy APC" through embedded APC function blocks, drag-and-drop configuration, pre-engineered applications and so on. It seems, however, that most industries (with the exception of large refineries and petrochem plants) do not effectively take advantage of this technology. As an APC practitioner for more than 20 years, I have seen dramatic shifts in ease-of-use for advanced control tools. Early APC projects were done with hard-coded Fortran programs running in a supervisory computer that took a programmer and a background in linear programming to keep running. PC's and the Microsoft era brought better user tools with easier data manipulation and graphicsl displays, but implementation still required a supervisory computer and all the peripheral data communications, historians, watch-dog timers, fail-shed logic and separate user interfaces. In today's environment, implementing a Model Predictive Controller or Neural Net soft sensor in the DeltaV is as simple as opening the APC Toolbar, dragging the APC block into a control module, wiring up the inputs, and right-clicking to launch the model identification tool -- all right in the DCS environment.
So, why aren't more industries, plants, engineers using these tools and getting the benefits from applying advanced process controls to their units?
I have some ideas. I'd be interested in yours.
Pete Sharpe, Emerson Process Management
- Login or register to post comments
- 3766 reads
- Permalink
- 2 comments
Its never that easy
I personally think "Easy APC" is what you end up with if you get some fresh marketing guys selling APC applications. If marketters sell APC to be something that just needs draggin' and droppin', then you know two things for sure i) they have no clue about the process flow of developing APC applications ii) they have no clue about the efforts and time spent in each of the development phase. What is the biggest workload of developing any APC application is CERTAINLY NOT the configuration that you do at the tail end, everything and i mean EVERYTHING is about understanding the process and all its quirkiness. Even the time spent to get to know the modelling tool pales in comparision to this, especially now that process modelling tools require less mathematical/statistical backgrounds to work on them. No, the biggest chunk remains the process and the weirdness almost everyone of them exhibits be it a simple two cut splitter to a complex discontinuous process such as delayed coker, each has its own uniqueness, even identical units in different trains side by side display differences. And its the appreciation and understanding of their fundamental character/nature/personality that is the bulk of the APC development work. For actually building the application (i.e. the dragging/dropping and connecting aka APC configuration phase), i tend to give a one day period (hell maybe two) for even complex APC apps. "Easy APC?????".... marketers, its all those damn marketers....
Regards,
Prakash K.K.
APC recognition hard -- even if implementation's "easy"
"Easy" loop-tuning has been around a lot longer than "easy" APC....yet by all accounts most plants out there don't yet have properly tuned PID loops. So, even if the tools are available, it seems that it takes a certain level of sophistication to even realize that a particular application would benefit from APC -- even if the benefits are but a mouse-click away.
Keith Larson